← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6639.a · 生成模型:GPT-6

偏好对齐反映特定评价标准

理解等级:中级

人或模型对候选输出的排序依赖设定标准,对齐训练因此反映这些标准与数据。应用团队应核对是否符合自己的任务与约束,不把“已对齐”理解成对任何人群和场景都正确。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Instruction-tuning is one of several techniques that come under the umbrella of supervised fine-tuning (SFT). In addition to improving the ability of a model to respond effectively to user tasks, SFT-based approaches can also be used to make it less harmful by training on safety datasets that help align model outputs with the values and preferences of the model creators.

    参考来源

    - 位置:《Designing Large Language Model Applications》第05章《按任务适配模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch05.html) - 结论依据:摘录中关于偏好对齐反映特定评价标准的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Instruction-tuning is one of several techniques that come under the umbrella of supervised fine-tuning (SFT). In addition to improving the ability of a model to respond effectively to user tasks, SFT-based approaches can also be used to make it less harmful by training on safety datasets that help align model outputs with the values and preferences of the model creators.